Microframes
Definition and structure#
A microframe reifies a binary relation as a frame, with two argument FEs: Domain and Range, named appropriately for that relation. The same relation can be reused across many lexical frames and conceptual schemas.
An ordinary lexical frame with two FEs is not automatically a microframe. Its modeling function must be a reusable relation component. A lexical frame describing possession or kinship is distinct from a microframe used to relate two participants in that scene.
Semantic categories#
Categories are multiple-valued tags, not frame namespaces. A microframe can carry several categories that describe different aspects of the relation. The documented category vocabulary is:
| Category | Meaning |
|---|---|
| Formal | Classification, taxonomic and other structural connections |
| Mereological | Parts, members, materials and composition |
| Spatial | Spatial or conventional-location connections |
| Quality | Qualities, values, measurements and parameters |
| Telic | Purpose and function |
| Agentive | Creation or agency |
| Causal | Causation without requiring an intentional agent |
| Experiential | Perception, emotion, cognition and sensation |
| Role | Conventional participation, activity, ability or use |
| Social | Institutional, normative, kinship and social connections |
| Semiotic | Aboutness and reference |
| Context | Relevant circumstances and contextualization |
| General | An insufficiently specified relation requiring refinement |
| Force | Manifested force relations |
| Topological | RCC-8-style connections, overlap, containment and equality |
| Cardinal | Relative spatial direction |
| Movement | Relative-motion relations |
| Functional | A relation treated as single-valued for a bearer in the relevant context |
For example, an agency relation may have Agentive, Causal and Force tags. Category overlap does not imply that the microframe has multiple namespaces.
Temporal ordering such as precedes differs from relative movement. Do not assume a Temporal category is available without checking the vocabulary; an unregistered category is not created by a documentation example.
The relation's DUL parent expresses its grounding. A GL quale can additionally record its formal, constitutive, telic or agentive characterization; this is distinct from the qualia labels on lexical FEs.
When a distinct microframe is justified#
Require a relation needed by the model and not already expressed by an existing microframe. Check synonyms, argument reversal, filler restrictions and compositions before proposing another identity.
- Reversing a symmetric relation does not create a new relation.
- An asymmetric converse can have its own identity and an inverse link.
- A filler-type restriction is not by itself a new relation.
- A composition should use existing relations where possible; a separately justified named composite needs explicit expansion into its components.
- Sharing a category does not establish semantic mutual exclusion. Any grouping used for exclusion needs an explicit, justified constraint, not inference from a browse tag.
These are modeling criteria, not claims that every stored microframe already satisfies them. See Conceptual microframes.
Structural relations between FEs#
rel_structure asserts that a microframe's relation holds between two FEs aligned to Domain and Range. The FEs can belong to one frame or different frames; distinguish those cases from a frame-to-frame claim.
This structure describes how participants are related. It does not assert that an entire lexical frame is the home of the relation's lexical meaning. Verify endpoints and mappings rather than guessing from participant names.
Lexical grounding through TQR#
rel_tqr relates a microframe to a lexical frame in which its relational content is expressed, with Domain/Range-to-FE correspondences beneath the frame link.
| Layer | What is asserted |
|---|---|
| Frame link | The lexical frame is a grounding home for the relation's content. |
| FE mappings | Domain and Range align with particular FEs of that lexical frame. |
| LU-level links | Relations among particular lexical readings, where recorded separately. |
This differs from rel_structure: one links lexical grounding, the other asserts a relation between particular FEs. A microframe can participate in both for different purposes. Neither Subframe nor Perspective_on is a substitute: their semantics concern component situations and viewpoints, not merely two related participants.
See TQR and TQR Reference. Keep frame links, FE alignments and LU links distinct during review.
Lexicalization and compatibility boundary#
Microframe reuse and TQR grounding do not prove that an LU should directly evoke the microframe. Direct microframe lexicalization versus lexical-frame grounding requires explicit policy review; do not infer a general permission or prohibition from binary arity.
Lemma types do not decide whether an LU may evoke a microframe (Lemmas §8). Review any existing lexicalizations on their actual meanings and modeling function, without automatic moves or deletion.
Relation to other ontological structures#
Microframes represent relations; they do not replace participant categories, filler restrictions or FE alternatives. An FE–Frame/LU–FE participant association has its own semantics and suitability requirement. CoreSet, Requires and Excludes constrain realization and must not be inferred from microframe category tags. See Frame elements and FrameNet patterns.